1 from __future__ import unicode_literals
16 from ..compat import (
19 compat_etree_fromstring,
25 compat_urllib_parse_unquote,
26 compat_urllib_parse_urlencode,
27 compat_urllib_request,
30 from ..downloader.f4m import remove_encrypted_media
54 parse_m3u8_attributes,
71 class InfoExtractor(object):
72 """Information Extractor class.
74 Information extractors are the classes that, given a URL, extract
75 information about the video (or videos) the URL refers to. This
76 information includes the real video URL, the video title, author and
77 others. The information is stored in a dictionary which is then
78 passed to the YoutubeDL. The YoutubeDL processes this
79 information possibly downloading the video to the file system, among
80 other possible outcomes.
82 The type field determines the type of the result.
83 By far the most common value (and the default if _type is missing) is
84 "video", which indicates a single video.
86 For a video, the dictionaries must include the following fields:
89 title: Video title, unescaped.
91 Additionally, it must contain either a formats entry or a url one:
93 formats: A list of dictionaries for each format available, ordered
94 from worst to best quality.
97 * url Mandatory. The URL of the video file
99 The URL of the manifest file in case of
100 fragmented media (DASH, hls, hds)
101 * ext Will be calculated from URL if missing
102 * format A human-readable description of the format
103 ("mp4 container with h264/opus").
104 Calculated from the format_id, width, height.
105 and format_note fields if missing.
106 * format_id A short description of the format
107 ("mp4_h264_opus" or "19").
108 Technically optional, but strongly recommended.
109 * format_note Additional info about the format
110 ("3D" or "DASH video")
111 * width Width of the video, if known
112 * height Height of the video, if known
113 * resolution Textual description of width and height
114 * tbr Average bitrate of audio and video in KBit/s
115 * abr Average audio bitrate in KBit/s
116 * acodec Name of the audio codec in use
117 * asr Audio sampling rate in Hertz
118 * vbr Average video bitrate in KBit/s
120 * vcodec Name of the video codec in use
121 * container Name of the container format
122 * filesize The number of bytes, if known in advance
123 * filesize_approx An estimate for the number of bytes
124 * player_url SWF Player URL (used for rtmpdump).
125 * protocol The protocol that will be used for the actual
126 download, lower-case.
127 "http", "https", "rtsp", "rtmp", "rtmpe",
128 "m3u8", "m3u8_native" or "http_dash_segments".
130 Base URL for fragments. Each fragment's path
131 value (if present) will be relative to
133 * fragments A list of fragments of a fragmented media.
134 Each fragment entry must contain either an url
135 or a path. If an url is present it should be
136 considered by a client. Otherwise both path and
137 fragment_base_url must be present. Here is
138 the list of all potential fields:
139 * "url" - fragment's URL
140 * "path" - fragment's path relative to
142 * "duration" (optional, int or float)
143 * "filesize" (optional, int)
144 * preference Order number of this format. If this field is
145 present and not None, the formats get sorted
146 by this field, regardless of all other values.
147 -1 for default (order by other properties),
148 -2 or smaller for less than default.
149 < -1000 to hide the format (if there is
150 another one which is strictly better)
151 * language Language code, e.g. "de" or "en-US".
152 * language_preference Is this in the language mentioned in
154 10 if it's what the URL is about,
155 -1 for default (don't know),
156 -10 otherwise, other values reserved for now.
157 * quality Order number of the video quality of this
158 format, irrespective of the file format.
159 -1 for default (order by other properties),
160 -2 or smaller for less than default.
161 * source_preference Order number for this video source
162 (quality takes higher priority)
163 -1 for default (order by other properties),
164 -2 or smaller for less than default.
165 * http_headers A dictionary of additional HTTP headers
166 to add to the request.
167 * stretched_ratio If given and not 1, indicates that the
168 video's pixels are not square.
169 width : height ratio as float.
170 * no_resume The server does not support resuming the
171 (HTTP or RTMP) download. Boolean.
173 url: Final video URL.
174 ext: Video filename extension.
175 format: The video format, defaults to ext (used for --get-format)
176 player_url: SWF Player URL (used for rtmpdump).
178 The following fields are optional:
180 alt_title: A secondary title of the video.
181 display_id An alternative identifier for the video, not necessarily
182 unique, but available before title. Typically, id is
183 something like "4234987", title "Dancing naked mole rats",
184 and display_id "dancing-naked-mole-rats"
185 thumbnails: A list of dictionaries, with the following entries:
186 * "id" (optional, string) - Thumbnail format ID
188 * "preference" (optional, int) - quality of the image
189 * "width" (optional, int)
190 * "height" (optional, int)
191 * "resolution" (optional, string "{width}x{height"},
193 * "filesize" (optional, int)
194 thumbnail: Full URL to a video thumbnail image.
195 description: Full video description.
196 uploader: Full name of the video uploader.
197 license: License name the video is licensed under.
198 creator: The creator of the video.
199 release_date: The date (YYYYMMDD) when the video was released.
200 timestamp: UNIX timestamp of the moment the video became available.
201 upload_date: Video upload date (YYYYMMDD).
202 If not explicitly set, calculated from timestamp.
203 uploader_id: Nickname or id of the video uploader.
204 uploader_url: Full URL to a personal webpage of the video uploader.
205 location: Physical location where the video was filmed.
206 subtitles: The available subtitles as a dictionary in the format
207 {tag: subformats}. "tag" is usually a language code, and
208 "subformats" is a list sorted from lower to higher
209 preference, each element is a dictionary with the "ext"
211 * "data": The subtitles file contents
212 * "url": A URL pointing to the subtitles file
213 "ext" will be calculated from URL if missing
214 automatic_captions: Like 'subtitles', used by the YoutubeIE for
215 automatically generated captions
216 duration: Length of the video in seconds, as an integer or float.
217 view_count: How many users have watched the video on the platform.
218 like_count: Number of positive ratings of the video
219 dislike_count: Number of negative ratings of the video
220 repost_count: Number of reposts of the video
221 average_rating: Average rating give by users, the scale used depends on the webpage
222 comment_count: Number of comments on the video
223 comments: A list of comments, each with one or more of the following
224 properties (all but one of text or html optional):
225 * "author" - human-readable name of the comment author
226 * "author_id" - user ID of the comment author
228 * "html" - Comment as HTML
229 * "text" - Plain text of the comment
230 * "timestamp" - UNIX timestamp of comment
231 * "parent" - ID of the comment this one is replying to.
232 Set to "root" to indicate that this is a
233 comment to the original video.
234 age_limit: Age restriction for the video, as an integer (years)
235 webpage_url: The URL to the video webpage, if given to youtube-dl it
236 should allow to get the same result again. (It will be set
237 by YoutubeDL if it's missing)
238 categories: A list of categories that the video falls in, for example
240 tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"]
241 is_live: True, False, or None (=unknown). Whether this video is a
242 live stream that goes on instead of a fixed-length video.
243 start_time: Time in seconds where the reproduction should start, as
244 specified in the URL.
245 end_time: Time in seconds where the reproduction should end, as
246 specified in the URL.
248 The following fields should only be used when the video belongs to some logical
251 chapter: Name or title of the chapter the video belongs to.
252 chapter_number: Number of the chapter the video belongs to, as an integer.
253 chapter_id: Id of the chapter the video belongs to, as a unicode string.
255 The following fields should only be used when the video is an episode of some
256 series, programme or podcast:
258 series: Title of the series or programme the video episode belongs to.
259 season: Title of the season the video episode belongs to.
260 season_number: Number of the season the video episode belongs to, as an integer.
261 season_id: Id of the season the video episode belongs to, as a unicode string.
262 episode: Title of the video episode. Unlike mandatory video title field,
263 this field should denote the exact title of the video episode
264 without any kind of decoration.
265 episode_number: Number of the video episode within a season, as an integer.
266 episode_id: Id of the video episode, as a unicode string.
268 The following fields should only be used when the media is a track or a part of
271 track: Title of the track.
272 track_number: Number of the track within an album or a disc, as an integer.
273 track_id: Id of the track (useful in case of custom indexing, e.g. 6.iii),
275 artist: Artist(s) of the track.
276 genre: Genre(s) of the track.
277 album: Title of the album the track belongs to.
278 album_type: Type of the album (e.g. "Demo", "Full-length", "Split", "Compilation", etc).
279 album_artist: List of all artists appeared on the album (e.g.
280 "Ash Borer / Fell Voices" or "Various Artists", useful for splits
282 disc_number: Number of the disc or other physical medium the track belongs to,
284 release_year: Year (YYYY) when the album was released.
286 Unless mentioned otherwise, the fields should be Unicode strings.
288 Unless mentioned otherwise, None is equivalent to absence of information.
291 _type "playlist" indicates multiple videos.
292 There must be a key "entries", which is a list, an iterable, or a PagedList
293 object, each element of which is a valid dictionary by this specification.
295 Additionally, playlists can have "title", "description" and "id" attributes
296 with the same semantics as videos (see above).
299 _type "multi_video" indicates that there are multiple videos that
300 form a single show, for examples multiple acts of an opera or TV episode.
301 It must have an entries key like a playlist and contain all the keys
302 required for a video at the same time.
305 _type "url" indicates that the video must be extracted from another
306 location, possibly by a different extractor. Its only required key is:
307 "url" - the next URL to extract.
308 The key "ie_key" can be set to the class name (minus the trailing "IE",
309 e.g. "Youtube") if the extractor class is known in advance.
310 Additionally, the dictionary may have any properties of the resolved entity
311 known in advance, for example "title" if the title of the referred video is
315 _type "url_transparent" entities have the same specification as "url", but
316 indicate that the given additional information is more precise than the one
317 associated with the resolved URL.
318 This is useful when a site employs a video service that hosts the video and
319 its technical metadata, but that video service does not embed a useful
320 title, description etc.
323 Subclasses of this one should re-define the _real_initialize() and
324 _real_extract() methods and define a _VALID_URL regexp.
325 Probably, they should also be added to the list of extractors.
327 _GEO_BYPASS attribute may be set to False in order to disable
328 geo restriction bypass mechanisms for a particular extractor.
329 Though it won't disable explicit geo restriction bypass based on
330 country code provided with geo_bypass_country. (experimental)
332 _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
333 countries for this extractor. One of these countries will be used by
334 geo restriction bypass mechanism right away in order to bypass
335 geo restriction, of course, if the mechanism is not disabled. (experimental)
337 NB: both these geo attributes are experimental and may change in future
338 or be completely removed.
340 Finally, the _WORKING attribute should be set to False for broken IEs
341 in order to warn the users and skip the tests.
346 _x_forwarded_for_ip = None
348 _GEO_COUNTRIES = None
351 def __init__(self, downloader=None):
352 """Constructor. Receives an optional downloader."""
354 self._x_forwarded_for_ip = None
355 self.set_downloader(downloader)
358 def suitable(cls, url):
359 """Receives a URL and returns True if suitable for this IE."""
361 # This does not use has/getattr intentionally - we want to know whether
362 # we have cached the regexp for *this* class, whereas getattr would also
363 # match the superclass
364 if '_VALID_URL_RE' not in cls.__dict__:
365 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
366 return cls._VALID_URL_RE.match(url) is not None
369 def _match_id(cls, url):
370 if '_VALID_URL_RE' not in cls.__dict__:
371 cls._VALID_URL_RE = re.compile(cls._VALID_URL)
372 m = cls._VALID_URL_RE.match(url)
378 """Getter method for _WORKING."""
381 def initialize(self):
382 """Initializes an instance (authentication, etc)."""
383 self._initialize_geo_bypass(self._GEO_COUNTRIES)
385 self._real_initialize()
388 def _initialize_geo_bypass(self, countries):
390 Initialize geo restriction bypass mechanism.
392 This method is used to initialize geo bypass mechanism based on faking
393 X-Forwarded-For HTTP header. A random country from provided country list
394 is selected and a random IP belonging to this country is generated. This
395 IP will be passed as X-Forwarded-For HTTP header in all subsequent
398 This method will be used for initial geo bypass mechanism initialization
399 during the instance initialization with _GEO_COUNTRIES.
401 You may also manually call it from extractor's code if geo countries
402 information is not available beforehand (e.g. obtained during
403 extraction) or due to some another reason.
405 if not self._x_forwarded_for_ip:
406 country_code = self._downloader.params.get('geo_bypass_country', None)
407 # If there is no explicit country for geo bypass specified and
408 # the extractor is known to be geo restricted let's fake IP
409 # as X-Forwarded-For right away.
410 if (not country_code and
412 self._downloader.params.get('geo_bypass', True) and
414 country_code = random.choice(countries)
416 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
417 if self._downloader.params.get('verbose', False):
418 self._downloader.to_stdout(
419 '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
420 % (self._x_forwarded_for_ip, country_code.upper()))
422 def extract(self, url):
423 """Extracts URL information and returns it in list of dicts."""
428 ie_result = self._real_extract(url)
429 if self._x_forwarded_for_ip:
430 ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
432 except GeoRestrictedError as e:
433 if self.__maybe_fake_ip_and_retry(e.countries):
436 except ExtractorError:
438 except compat_http_client.IncompleteRead as e:
439 raise ExtractorError('A network error has occurred.', cause=e, expected=True)
440 except (KeyError, StopIteration) as e:
441 raise ExtractorError('An extractor error has occurred.', cause=e)
443 def __maybe_fake_ip_and_retry(self, countries):
444 if (not self._downloader.params.get('geo_bypass_country', None) and
446 self._downloader.params.get('geo_bypass', True) and
447 not self._x_forwarded_for_ip and
449 country_code = random.choice(countries)
450 self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
451 if self._x_forwarded_for_ip:
453 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
454 % (self._x_forwarded_for_ip, country_code.upper()))
458 def set_downloader(self, downloader):
459 """Sets the downloader for this IE."""
460 self._downloader = downloader
462 def _real_initialize(self):
463 """Real initialization process. Redefine in subclasses."""
466 def _real_extract(self, url):
467 """Real extraction process. Redefine in subclasses."""
472 """A string for getting the InfoExtractor with get_info_extractor"""
473 return compat_str(cls.__name__[:-2])
477 return compat_str(type(self).__name__[:-2])
479 def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
480 """ Returns the response handle """
482 self.report_download_webpage(video_id)
483 elif note is not False:
485 self.to_screen('%s' % (note,))
487 self.to_screen('%s: %s' % (video_id, note))
488 if isinstance(url_or_request, compat_urllib_request.Request):
489 url_or_request = update_Request(
490 url_or_request, data=data, headers=headers, query=query)
493 url_or_request = update_url_query(url_or_request, query)
494 if data is not None or headers:
495 url_or_request = sanitized_Request(url_or_request, data, headers)
497 return self._downloader.urlopen(url_or_request)
498 except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
502 errnote = 'Unable to download webpage'
504 errmsg = '%s: %s' % (errnote, error_to_compat_str(err))
506 raise ExtractorError(errmsg, sys.exc_info()[2], cause=err)
508 self._downloader.report_warning(errmsg)
511 def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}):
512 """ Returns a tuple (page content as string, URL handle) """
513 # Strip hashes from the URL (#1038)
514 if isinstance(url_or_request, (compat_str, str)):
515 url_or_request = url_or_request.partition('#')[0]
517 # Some sites check X-Forwarded-For HTTP header in order to figure out
518 # the origin of the client behind proxy. This allows bypassing geo
519 # restriction by faking this header's value to IP that belongs to some
520 # geo unrestricted country. We will do so once we encounter any
521 # geo restriction error.
522 if self._x_forwarded_for_ip:
523 if 'X-Forwarded-For' not in headers:
524 headers['X-Forwarded-For'] = self._x_forwarded_for_ip
526 urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
530 content = self._webpage_read_content(urlh, url_or_request, video_id, note, errnote, fatal, encoding=encoding)
531 return (content, urlh)
534 def _guess_encoding_from_content(content_type, webpage_bytes):
535 m = re.match(r'[a-zA-Z0-9_.-]+/[a-zA-Z0-9_.-]+\s*;\s*charset=(.+)', content_type)
537 encoding = m.group(1)
539 m = re.search(br'<meta[^>]+charset=[\'"]?([^\'")]+)[ /\'">]',
540 webpage_bytes[:1024])
542 encoding = m.group(1).decode('ascii')
543 elif webpage_bytes.startswith(b'\xff\xfe'):
550 def __check_blocked(self, content):
551 first_block = content[:512]
552 if ('<title>Access to this site is blocked</title>' in content and
553 'Websense' in first_block):
554 msg = 'Access to this webpage has been blocked by Websense filtering software in your network.'
555 blocked_iframe = self._html_search_regex(
556 r'<iframe src="([^"]+)"', content,
557 'Websense information URL', default=None)
559 msg += ' Visit %s for more details' % blocked_iframe
560 raise ExtractorError(msg, expected=True)
561 if '<title>The URL you requested has been blocked</title>' in first_block:
563 'Access to this webpage has been blocked by Indian censorship. '
564 'Use a VPN or proxy server (with --proxy) to route around it.')
565 block_msg = self._html_search_regex(
566 r'</h1><p>(.*?)</p>',
567 content, 'block message', default=None)
569 msg += ' (Message: "%s")' % block_msg.replace('\n', ' ')
570 raise ExtractorError(msg, expected=True)
571 if ('<title>TTK :: Доступ к ресурсу ограничен</title>' in content and
572 'blocklist.rkn.gov.ru' in content):
573 raise ExtractorError(
574 'Access to this webpage has been blocked by decision of the Russian government. '
575 'Visit http://blocklist.rkn.gov.ru/ for a block reason.',
578 def _webpage_read_content(self, urlh, url_or_request, video_id, note=None, errnote=None, fatal=True, prefix=None, encoding=None):
579 content_type = urlh.headers.get('Content-Type', '')
580 webpage_bytes = urlh.read()
581 if prefix is not None:
582 webpage_bytes = prefix + webpage_bytes
584 encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
585 if self._downloader.params.get('dump_intermediate_pages', False):
587 url = url_or_request.get_full_url()
588 except AttributeError:
590 self.to_screen('Dumping request to ' + url)
591 dump = base64.b64encode(webpage_bytes).decode('ascii')
592 self._downloader.to_screen(dump)
593 if self._downloader.params.get('write_pages', False):
595 url = url_or_request.get_full_url()
596 except AttributeError:
598 basen = '%s_%s' % (video_id, url)
600 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
601 basen = basen[:240 - len(h)] + h
602 raw_filename = basen + '.dump'
603 filename = sanitize_filename(raw_filename, restricted=True)
604 self.to_screen('Saving request to ' + filename)
605 # Working around MAX_PATH limitation on Windows (see
606 # http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx)
607 if compat_os_name == 'nt':
608 absfilepath = os.path.abspath(filename)
609 if len(absfilepath) > 259:
610 filename = '\\\\?\\' + absfilepath
611 with open(filename, 'wb') as outf:
612 outf.write(webpage_bytes)
615 content = webpage_bytes.decode(encoding, 'replace')
617 content = webpage_bytes.decode('utf-8', 'replace')
619 self.__check_blocked(content)
623 def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None, data=None, headers={}, query={}):
624 """ Returns the data of the page as a string """
627 while success is False:
629 res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query)
631 except compat_http_client.IncompleteRead as e:
633 if try_count >= tries:
635 self._sleep(timeout, video_id)
642 def _download_xml(self, url_or_request, video_id,
643 note='Downloading XML', errnote='Unable to download XML',
644 transform_source=None, fatal=True, encoding=None, data=None, headers={}, query={}):
645 """Return the xml as an xml.etree.ElementTree.Element"""
646 xml_string = self._download_webpage(
647 url_or_request, video_id, note, errnote, fatal=fatal, encoding=encoding, data=data, headers=headers, query=query)
648 if xml_string is False:
651 xml_string = transform_source(xml_string)
652 return compat_etree_fromstring(xml_string.encode('utf-8'))
654 def _download_json(self, url_or_request, video_id,
655 note='Downloading JSON metadata',
656 errnote='Unable to download JSON metadata',
657 transform_source=None,
658 fatal=True, encoding=None, data=None, headers={}, query={}):
659 json_string = self._download_webpage(
660 url_or_request, video_id, note, errnote, fatal=fatal,
661 encoding=encoding, data=data, headers=headers, query=query)
662 if (not fatal) and json_string is False:
664 return self._parse_json(
665 json_string, video_id, transform_source=transform_source, fatal=fatal)
667 def _parse_json(self, json_string, video_id, transform_source=None, fatal=True):
669 json_string = transform_source(json_string)
671 return json.loads(json_string)
672 except ValueError as ve:
673 errmsg = '%s: Failed to parse JSON ' % video_id
675 raise ExtractorError(errmsg, cause=ve)
677 self.report_warning(errmsg + str(ve))
679 def report_warning(self, msg, video_id=None):
680 idstr = '' if video_id is None else '%s: ' % video_id
681 self._downloader.report_warning(
682 '[%s] %s%s' % (self.IE_NAME, idstr, msg))
684 def to_screen(self, msg):
685 """Print msg to screen, prefixing it with '[ie_name]'"""
686 self._downloader.to_screen('[%s] %s' % (self.IE_NAME, msg))
688 def report_extraction(self, id_or_name):
689 """Report information extraction."""
690 self.to_screen('%s: Extracting information' % id_or_name)
692 def report_download_webpage(self, video_id):
693 """Report webpage download."""
694 self.to_screen('%s: Downloading webpage' % video_id)
696 def report_age_confirmation(self):
697 """Report attempt to confirm age."""
698 self.to_screen('Confirming age')
700 def report_login(self):
701 """Report attempt to log in."""
702 self.to_screen('Logging in')
705 def raise_login_required(msg='This video is only available for registered users'):
706 raise ExtractorError(
707 '%s. Use --username and --password or --netrc to provide account credentials.' % msg,
711 def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
712 raise GeoRestrictedError(msg, countries=countries)
714 # Methods for following #608
716 def url_result(url, ie=None, video_id=None, video_title=None):
717 """Returns a URL that points to a page that should be processed"""
718 # TODO: ie should be the class used for getting the info
719 video_info = {'_type': 'url',
722 if video_id is not None:
723 video_info['id'] = video_id
724 if video_title is not None:
725 video_info['title'] = video_title
728 def playlist_from_matches(self, matches, video_id, video_title, getter=None, ie=None):
730 self.url_result(self._proto_relative_url(getter(m) if getter else m), ie)
732 return self.playlist_result(
733 urlrs, playlist_id=video_id, playlist_title=video_title)
736 def playlist_result(entries, playlist_id=None, playlist_title=None, playlist_description=None):
737 """Returns a playlist"""
738 video_info = {'_type': 'playlist',
741 video_info['id'] = playlist_id
743 video_info['title'] = playlist_title
744 if playlist_description:
745 video_info['description'] = playlist_description
748 def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
750 Perform a regex search on the given string, using a single or a list of
751 patterns returning the first matching group.
752 In case of failure return a default value or raise a WARNING or a
753 RegexNotFoundError, depending on fatal, specifying the field name.
755 if isinstance(pattern, (str, compat_str, compiled_regex_type)):
756 mobj = re.search(pattern, string, flags)
759 mobj = re.search(p, string, flags)
763 if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
764 _name = '\033[0;34m%s\033[0m' % name
770 # return the first matching group
771 return next(g for g in mobj.groups() if g is not None)
773 return mobj.group(group)
774 elif default is not NO_DEFAULT:
777 raise RegexNotFoundError('Unable to extract %s' % _name)
779 self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message())
782 def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
784 Like _search_regex, but strips HTML tags and unescapes entities.
786 res = self._search_regex(pattern, string, name, default, fatal, flags, group)
788 return clean_html(res).strip()
792 def _get_netrc_login_info(self, netrc_machine=None):
795 netrc_machine = netrc_machine or self._NETRC_MACHINE
797 if self._downloader.params.get('usenetrc', False):
799 info = netrc.netrc().authenticators(netrc_machine)
804 raise netrc.NetrcParseError(
805 'No authenticators for %s' % netrc_machine)
806 except (IOError, netrc.NetrcParseError) as err:
807 self._downloader.report_warning(
808 'parsing .netrc: %s' % error_to_compat_str(err))
810 return username, password
812 def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
814 Get the login info as (username, password)
815 First look for the manually specified credentials using username_option
816 and password_option as keys in params dictionary. If no such credentials
817 available look in the netrc file using the netrc_machine or _NETRC_MACHINE
819 If there's no info available, return (None, None)
821 if self._downloader is None:
824 downloader_params = self._downloader.params
826 # Attempt to use provided username and password or .netrc data
827 if downloader_params.get(username_option) is not None:
828 username = downloader_params[username_option]
829 password = downloader_params[password_option]
831 username, password = self._get_netrc_login_info(netrc_machine)
833 return username, password
835 def _get_tfa_info(self, note='two-factor verification code'):
837 Get the two-factor authentication info
838 TODO - asking the user will be required for sms/phone verify
839 currently just uses the command line option
840 If there's no info available, return None
842 if self._downloader is None:
844 downloader_params = self._downloader.params
846 if downloader_params.get('twofactor') is not None:
847 return downloader_params['twofactor']
849 return compat_getpass('Type %s and press [Return]: ' % note)
851 # Helper functions for extracting OpenGraph info
853 def _og_regexes(prop):
854 content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
855 property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
856 % {'prop': re.escape(prop)})
857 template = r'<meta[^>]+?%s[^>]+?%s'
859 template % (property_re, content_re),
860 template % (content_re, property_re),
864 def _meta_regex(prop):
865 return r'''(?isx)<meta
866 (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1)
867 [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop)
869 def _og_search_property(self, prop, html, name=None, **kargs):
870 if not isinstance(prop, (list, tuple)):
873 name = 'OpenGraph %s' % prop[0]
876 og_regexes.extend(self._og_regexes(p))
877 escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs)
880 return unescapeHTML(escaped)
882 def _og_search_thumbnail(self, html, **kargs):
883 return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs)
885 def _og_search_description(self, html, **kargs):
886 return self._og_search_property('description', html, fatal=False, **kargs)
888 def _og_search_title(self, html, **kargs):
889 return self._og_search_property('title', html, **kargs)
891 def _og_search_video_url(self, html, name='video url', secure=True, **kargs):
892 regexes = self._og_regexes('video') + self._og_regexes('video:url')
894 regexes = self._og_regexes('video:secure_url') + regexes
895 return self._html_search_regex(regexes, html, name, **kargs)
897 def _og_search_url(self, html, **kargs):
898 return self._og_search_property('url', html, **kargs)
900 def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
901 if not isinstance(name, (list, tuple)):
903 if display_name is None:
904 display_name = name[0]
905 return self._html_search_regex(
906 [self._meta_regex(n) for n in name],
907 html, display_name, fatal=fatal, group='content', **kwargs)
909 def _dc_search_uploader(self, html):
910 return self._html_search_meta('dc.creator', html, 'uploader')
912 def _rta_search(self, html):
913 # See http://www.rtalabel.org/index.php?content=howtofaq#single
914 if re.search(r'(?ix)<meta\s+name="rating"\s+'
915 r' content="RTA-5042-1996-1400-1577-RTA"',
920 def _media_rating_search(self, html):
921 # See http://www.tjg-designs.com/WP/metadata-code-examples-adding-metadata-to-your-web-pages/
922 rating = self._html_search_meta('rating', html)
934 return RATING_TABLE.get(rating.lower())
936 def _family_friendly_search(self, html):
937 # See http://schema.org/VideoObject
938 family_friendly = self._html_search_meta('isFamilyFriendly', html)
940 if not family_friendly:
949 return RATING_TABLE.get(family_friendly.lower())
951 def _twitter_search_player(self, html):
952 return self._html_search_meta('twitter:player', html,
953 'twitter card player')
955 def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
956 json_ld = self._search_regex(
957 r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
958 html, 'JSON-LD', group='json_ld', **kwargs)
959 default = kwargs.get('default', NO_DEFAULT)
961 return default if default is not NO_DEFAULT else {}
962 # JSON-LD may be malformed and thus `fatal` should be respected.
963 # At the same time `default` may be passed that assumes `fatal=False`
964 # for _search_regex. Let's simulate the same behavior here as well.
965 fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
966 return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
968 def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
969 if isinstance(json_ld, compat_str):
970 json_ld = self._parse_json(json_ld, video_id, fatal=fatal)
974 if not isinstance(json_ld, (list, tuple, dict)):
976 if isinstance(json_ld, dict):
979 if e.get('@context') == 'http://schema.org':
980 item_type = e.get('@type')
981 if expected_type is not None and expected_type != item_type:
983 if item_type == 'TVEpisode':
985 'episode': unescapeHTML(e.get('name')),
986 'episode_number': int_or_none(e.get('episodeNumber')),
987 'description': unescapeHTML(e.get('description')),
989 part_of_season = e.get('partOfSeason')
990 if isinstance(part_of_season, dict) and part_of_season.get('@type') == 'TVSeason':
991 info['season_number'] = int_or_none(part_of_season.get('seasonNumber'))
992 part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
993 if isinstance(part_of_series, dict) and part_of_series.get('@type') == 'TVSeries':
994 info['series'] = unescapeHTML(part_of_series.get('name'))
995 elif item_type == 'Article':
997 'timestamp': parse_iso8601(e.get('datePublished')),
998 'title': unescapeHTML(e.get('headline')),
999 'description': unescapeHTML(e.get('articleBody')),
1001 elif item_type == 'VideoObject':
1003 'url': e.get('contentUrl'),
1004 'title': unescapeHTML(e.get('name')),
1005 'description': unescapeHTML(e.get('description')),
1006 'thumbnail': e.get('thumbnailUrl') or e.get('thumbnailURL'),
1007 'duration': parse_duration(e.get('duration')),
1008 'timestamp': unified_timestamp(e.get('uploadDate')),
1009 'filesize': float_or_none(e.get('contentSize')),
1010 'tbr': int_or_none(e.get('bitrate')),
1011 'width': int_or_none(e.get('width')),
1012 'height': int_or_none(e.get('height')),
1015 return dict((k, v) for k, v in info.items() if v is not None)
1018 def _hidden_inputs(html):
1019 html = re.sub(r'<!--(?:(?!<!--).)*-->', '', html)
1021 for input in re.findall(r'(?i)(<input[^>]+>)', html):
1022 attrs = extract_attributes(input)
1025 if attrs.get('type') not in ('hidden', 'submit'):
1027 name = attrs.get('name') or attrs.get('id')
1028 value = attrs.get('value')
1029 if name and value is not None:
1030 hidden_inputs[name] = value
1031 return hidden_inputs
1033 def _form_hidden_inputs(self, form_id, html):
1034 form = self._search_regex(
1035 r'(?is)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id,
1036 html, '%s form' % form_id, group='form')
1037 return self._hidden_inputs(form)
1039 def _sort_formats(self, formats, field_preference=None):
1041 raise ExtractorError('No video formats found')
1044 # Automatically determine tbr when missing based on abr and vbr (improves
1045 # formats sorting in some cases)
1046 if 'tbr' not in f and f.get('abr') is not None and f.get('vbr') is not None:
1047 f['tbr'] = f['abr'] + f['vbr']
1049 def _formats_key(f):
1050 # TODO remove the following workaround
1051 from ..utils import determine_ext
1052 if not f.get('ext') and 'url' in f:
1053 f['ext'] = determine_ext(f['url'])
1055 if isinstance(field_preference, (list, tuple)):
1058 if f.get(field) is not None
1059 else ('' if field == 'format_id' else -1)
1060 for field in field_preference)
1062 preference = f.get('preference')
1063 if preference is None:
1065 if f.get('ext') in ['f4f', 'f4m']: # Not yet supported
1068 protocol = f.get('protocol') or determine_protocol(f)
1069 proto_preference = 0 if protocol in ['http', 'https'] else (-0.5 if protocol == 'rtsp' else -0.1)
1071 if f.get('vcodec') == 'none': # audio only
1073 if self._downloader.params.get('prefer_free_formats'):
1074 ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
1076 ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a']
1079 audio_ext_preference = ORDER.index(f['ext'])
1081 audio_ext_preference = -1
1083 if f.get('acodec') == 'none': # video only
1085 if self._downloader.params.get('prefer_free_formats'):
1086 ORDER = ['flv', 'mp4', 'webm']
1088 ORDER = ['webm', 'flv', 'mp4']
1090 ext_preference = ORDER.index(f['ext'])
1093 audio_ext_preference = 0
1097 f.get('language_preference') if f.get('language_preference') is not None else -1,
1098 f.get('quality') if f.get('quality') is not None else -1,
1099 f.get('tbr') if f.get('tbr') is not None else -1,
1100 f.get('filesize') if f.get('filesize') is not None else -1,
1101 f.get('vbr') if f.get('vbr') is not None else -1,
1102 f.get('height') if f.get('height') is not None else -1,
1103 f.get('width') if f.get('width') is not None else -1,
1106 f.get('abr') if f.get('abr') is not None else -1,
1107 audio_ext_preference,
1108 f.get('fps') if f.get('fps') is not None else -1,
1109 f.get('filesize_approx') if f.get('filesize_approx') is not None else -1,
1110 f.get('source_preference') if f.get('source_preference') is not None else -1,
1111 f.get('format_id') if f.get('format_id') is not None else '',
1113 formats.sort(key=_formats_key)
1115 def _check_formats(self, formats, video_id):
1117 formats[:] = filter(
1118 lambda f: self._is_valid_url(
1120 item='%s video format' % f.get('format_id') if f.get('format_id') else 'video'),
1124 def _remove_duplicate_formats(formats):
1128 if f['url'] not in format_urls:
1129 format_urls.add(f['url'])
1130 unique_formats.append(f)
1131 formats[:] = unique_formats
1133 def _is_valid_url(self, url, video_id, item='video', headers={}):
1134 url = self._proto_relative_url(url, scheme='http:')
1135 # For now assume non HTTP(S) URLs always valid
1136 if not (url.startswith('http://') or url.startswith('https://')):
1139 self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
1141 except ExtractorError as e:
1142 if isinstance(e.cause, compat_urllib_error.URLError):
1144 '%s: %s URL is invalid, skipping' % (video_id, item))
1148 def http_scheme(self):
1149 """ Either "http:" or "https:", depending on the user's preferences """
1152 if self._downloader.params.get('prefer_insecure', False)
1155 def _proto_relative_url(self, url, scheme=None):
1158 if url.startswith('//'):
1160 scheme = self.http_scheme()
1165 def _sleep(self, timeout, video_id, msg_template=None):
1166 if msg_template is None:
1167 msg_template = '%(video_id)s: Waiting for %(timeout)s seconds'
1168 msg = msg_template % {'video_id': video_id, 'timeout': timeout}
1172 def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None,
1173 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1174 fatal=True, m3u8_id=None):
1175 manifest = self._download_xml(
1176 manifest_url, video_id, 'Downloading f4m manifest',
1177 'Unable to download f4m manifest',
1178 # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
1179 # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
1180 transform_source=transform_source,
1183 if manifest is False:
1186 return self._parse_f4m_formats(
1187 manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1188 transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id)
1190 def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
1191 transform_source=lambda s: fix_xml_ampersands(s).strip(),
1192 fatal=True, m3u8_id=None):
1193 # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy
1194 akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0')
1195 if akamai_pv is not None and ';' in akamai_pv.text:
1196 playerVerificationChallenge = akamai_pv.text.split(';')[0]
1197 if playerVerificationChallenge.strip() != '':
1201 manifest_version = '1.0'
1202 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
1204 manifest_version = '2.0'
1205 media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
1206 # Remove unsupported DRM protected media from final formats
1207 # rendition (see https://github.com/rg3/youtube-dl/issues/8573).
1208 media_nodes = remove_encrypted_media(media_nodes)
1211 base_url = xpath_text(
1212 manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
1213 'base URL', default=None)
1215 base_url = base_url.strip()
1217 bootstrap_info = xpath_element(
1218 manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
1219 'bootstrap info', default=None)
1222 mime_type = xpath_text(
1223 manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
1224 'base URL', default=None)
1225 if mime_type and mime_type.startswith('audio/'):
1228 for i, media_el in enumerate(media_nodes):
1229 tbr = int_or_none(media_el.attrib.get('bitrate'))
1230 width = int_or_none(media_el.attrib.get('width'))
1231 height = int_or_none(media_el.attrib.get('height'))
1232 format_id = '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)]))
1233 # If <bootstrapInfo> is present, the specified f4m is a
1234 # stream-level manifest, and only set-level manifests may refer to
1235 # external resources. See section 11.4 and section 4 of F4M spec
1236 if bootstrap_info is None:
1238 # @href is introduced in 2.0, see section 11.6 of F4M spec
1239 if manifest_version == '2.0':
1240 media_url = media_el.attrib.get('href')
1241 if media_url is None:
1242 media_url = media_el.attrib.get('url')
1246 media_url if media_url.startswith('http://') or media_url.startswith('https://')
1247 else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
1248 # If media_url is itself a f4m manifest do the recursive extraction
1249 # since bitrates in parent manifest (this one) and media_url manifest
1250 # may differ leading to inability to resolve the format by requested
1251 # bitrate in f4m downloader
1252 ext = determine_ext(manifest_url)
1254 f4m_formats = self._extract_f4m_formats(
1255 manifest_url, video_id, preference=preference, f4m_id=f4m_id,
1256 transform_source=transform_source, fatal=fatal)
1257 # Sometimes stream-level manifest contains single media entry that
1258 # does not contain any quality metadata (e.g. http://matchtv.ru/#live-player).
1259 # At the same time parent's media entry in set-level manifest may
1260 # contain it. We will copy it from parent in such cases.
1261 if len(f4m_formats) == 1:
1264 'tbr': f.get('tbr') or tbr,
1265 'width': f.get('width') or width,
1266 'height': f.get('height') or height,
1267 'format_id': f.get('format_id') if not tbr else format_id,
1270 formats.extend(f4m_formats)
1273 formats.extend(self._extract_m3u8_formats(
1274 manifest_url, video_id, 'mp4', preference=preference,
1275 m3u8_id=m3u8_id, fatal=fatal))
1278 'format_id': format_id,
1279 'url': manifest_url,
1280 'manifest_url': manifest_url,
1281 'ext': 'flv' if bootstrap_info is not None else None,
1286 'preference': preference,
1290 def _m3u8_meta_format(self, m3u8_url, ext=None, preference=None, m3u8_id=None):
1292 'format_id': '-'.join(filter(None, [m3u8_id, 'meta'])),
1296 'preference': preference - 100 if preference else -100,
1297 'resolution': 'multiple',
1298 'format_note': 'Quality selection URL',
1301 def _extract_m3u8_formats(self, m3u8_url, video_id, ext=None,
1302 entry_protocol='m3u8', preference=None,
1303 m3u8_id=None, note=None, errnote=None,
1304 fatal=True, live=False):
1306 res = self._download_webpage_handle(
1308 note=note or 'Downloading m3u8 information',
1309 errnote=errnote or 'Failed to download m3u8 information',
1313 m3u8_doc, urlh = res
1314 m3u8_url = urlh.geturl()
1316 if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
1319 formats = [self._m3u8_meta_format(m3u8_url, ext, preference, m3u8_id)]
1321 format_url = lambda u: (
1323 if re.match(r'^https?://', u)
1324 else compat_urlparse.urljoin(m3u8_url, u))
1326 # We should try extracting formats only from master playlists [1], i.e.
1327 # playlists that describe available qualities. On the other hand media
1328 # playlists [2] should be returned as is since they contain just the media
1329 # without qualities renditions.
1330 # Fortunately, master playlist can be easily distinguished from media
1331 # playlist based on particular tags availability. As of [1, 2] master
1332 # playlist tags MUST NOT appear in a media playist and vice versa.
1333 # As of [3] #EXT-X-TARGETDURATION tag is REQUIRED for every media playlist
1334 # and MUST NOT appear in master playlist thus we can clearly detect media
1335 # playlist with this criterion.
1336 # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.4
1337 # 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3
1338 # 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.1
1339 if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is
1342 'format_id': m3u8_id,
1344 'protocol': entry_protocol,
1345 'preference': preference,
1347 audio_in_video_stream = {}
1350 for line in m3u8_doc.splitlines():
1351 if line.startswith('#EXT-X-STREAM-INF:'):
1352 last_info = parse_m3u8_attributes(line)
1353 elif line.startswith('#EXT-X-MEDIA:'):
1354 media = parse_m3u8_attributes(line)
1355 media_type = media.get('TYPE')
1356 if media_type in ('VIDEO', 'AUDIO'):
1357 group_id = media.get('GROUP-ID')
1358 media_url = media.get('URI')
1361 for v in (group_id, media.get('NAME')):
1365 'format_id': '-'.join(format_id),
1366 'url': format_url(media_url),
1367 'language': media.get('LANGUAGE'),
1369 'protocol': entry_protocol,
1370 'preference': preference,
1372 if media_type == 'AUDIO':
1373 f['vcodec'] = 'none'
1374 if group_id and not audio_in_video_stream.get(group_id):
1375 audio_in_video_stream[group_id] = False
1378 # When there is no URI in EXT-X-MEDIA let this tag's
1379 # data be used by regular URI lines below
1381 if media_type == 'AUDIO' and group_id:
1382 audio_in_video_stream[group_id] = True
1383 elif line.startswith('#') or not line.strip():
1386 tbr = int_or_none(last_info.get('AVERAGE-BANDWIDTH') or last_info.get('BANDWIDTH'), scale=1000)
1389 format_id.append(m3u8_id)
1390 # Despite specification does not mention NAME attribute for
1391 # EXT-X-STREAM-INF it still sometimes may be present
1392 stream_name = last_info.get('NAME') or last_media.get('NAME')
1393 # Bandwidth of live streams may differ over time thus making
1394 # format_id unpredictable. So it's better to keep provided
1397 format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
1398 manifest_url = format_url(line.strip())
1400 'format_id': '-'.join(format_id),
1401 'url': manifest_url,
1402 'manifest_url': manifest_url,
1405 'fps': float_or_none(last_info.get('FRAME-RATE')),
1406 'protocol': entry_protocol,
1407 'preference': preference,
1409 resolution = last_info.get('RESOLUTION')
1411 mobj = re.search(r'(?P<width>\d+)[xX](?P<height>\d+)', resolution)
1413 f['width'] = int(mobj.group('width'))
1414 f['height'] = int(mobj.group('height'))
1415 # Unified Streaming Platform
1417 r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
1419 abr, vbr = mobj.groups()
1420 abr, vbr = float_or_none(abr, 1000), float_or_none(vbr, 1000)
1425 f.update(parse_codecs(last_info.get('CODECS')))
1426 if audio_in_video_stream.get(last_info.get('AUDIO')) is False and f['vcodec'] != 'none':
1427 # TODO: update acodec for audio only formats with the same GROUP-ID
1428 f['acodec'] = 'none'
1435 def _xpath_ns(path, namespace=None):
1439 for c in path.split('/'):
1440 if not c or c == '.':
1443 out.append('{%s}%s' % (namespace, c))
1444 return '/'.join(out)
1446 def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
1447 smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
1453 namespace = self._parse_smil_namespace(smil)
1455 return self._parse_smil_formats(
1456 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1458 def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None):
1459 smil = self._download_smil(smil_url, video_id, fatal=fatal)
1462 return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
1464 def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
1465 return self._download_xml(
1466 smil_url, video_id, 'Downloading SMIL file',
1467 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
1469 def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
1470 namespace = self._parse_smil_namespace(smil)
1472 formats = self._parse_smil_formats(
1473 smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params)
1474 subtitles = self._parse_smil_subtitles(smil, namespace=namespace)
1476 video_id = os.path.splitext(url_basename(smil_url))[0]
1480 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1481 name = meta.attrib.get('name')
1482 content = meta.attrib.get('content')
1483 if not name or not content:
1485 if not title and name == 'title':
1487 elif not description and name in ('description', 'abstract'):
1488 description = content
1489 elif not upload_date and name == 'date':
1490 upload_date = unified_strdate(content)
1493 'id': image.get('type'),
1494 'url': image.get('src'),
1495 'width': int_or_none(image.get('width')),
1496 'height': int_or_none(image.get('height')),
1497 } for image in smil.findall(self._xpath_ns('.//image', namespace)) if image.get('src')]
1501 'title': title or video_id,
1502 'description': description,
1503 'upload_date': upload_date,
1504 'thumbnails': thumbnails,
1506 'subtitles': subtitles,
1509 def _parse_smil_namespace(self, smil):
1510 return self._search_regex(
1511 r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None)
1513 def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None, transform_rtmp_url=None):
1515 for meta in smil.findall(self._xpath_ns('./head/meta', namespace)):
1516 b = meta.get('base') or meta.get('httpBase')
1527 media = smil.findall(self._xpath_ns('.//video', namespace)) + smil.findall(self._xpath_ns('.//audio', namespace))
1528 for medium in media:
1529 src = medium.get('src')
1530 if not src or src in srcs:
1534 bitrate = float_or_none(medium.get('system-bitrate') or medium.get('systemBitrate'), 1000)
1535 filesize = int_or_none(medium.get('size') or medium.get('fileSize'))
1536 width = int_or_none(medium.get('width'))
1537 height = int_or_none(medium.get('height'))
1538 proto = medium.get('proto')
1539 ext = medium.get('ext')
1540 src_ext = determine_ext(src)
1541 streamer = medium.get('streamer') or base
1543 if proto == 'rtmp' or streamer.startswith('rtmp'):
1549 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate),
1551 'filesize': filesize,
1555 if transform_rtmp_url:
1556 streamer, src = transform_rtmp_url(streamer, src)
1557 formats[-1].update({
1563 src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src)
1564 src_url = src_url.strip()
1566 if proto == 'm3u8' or src_ext == 'm3u8':
1567 m3u8_formats = self._extract_m3u8_formats(
1568 src_url, video_id, ext or 'mp4', m3u8_id='hls', fatal=False)
1569 if len(m3u8_formats) == 1:
1571 m3u8_formats[0].update({
1572 'format_id': 'hls-%d' % (m3u8_count if bitrate is None else bitrate),
1577 formats.extend(m3u8_formats)
1580 if src_ext == 'f4m':
1585 'plugin': 'flowplayer-3.2.0.1',
1587 f4m_url += '&' if '?' in f4m_url else '?'
1588 f4m_url += compat_urllib_parse_urlencode(f4m_params)
1589 formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds', fatal=False))
1592 if src_url.startswith('http') and self._is_valid_url(src, video_id):
1596 'ext': ext or src_ext or 'flv',
1597 'format_id': 'http-%d' % (bitrate or http_count),
1599 'filesize': filesize,
1607 def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'):
1610 for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))):
1611 src = textstream.get('src')
1612 if not src or src in urls:
1615 ext = textstream.get('ext') or mimetype2ext(textstream.get('type')) or determine_ext(src)
1616 lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') or textstream.get('lang') or subtitles_lang
1617 subtitles.setdefault(lang, []).append({
1623 def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
1624 xspf = self._download_xml(
1625 playlist_url, playlist_id, 'Downloading xpsf playlist',
1626 'Unable to download xspf manifest', fatal=fatal)
1629 return self._parse_xspf(xspf, playlist_id)
1631 def _parse_xspf(self, playlist, playlist_id):
1633 'xspf': 'http://xspf.org/ns/0/',
1634 's1': 'http://static.streamone.nl/player/ns/0',
1638 for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
1640 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
1641 description = xpath_text(
1642 track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description')
1643 thumbnail = xpath_text(
1644 track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail')
1645 duration = float_or_none(
1646 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
1649 'url': location.text,
1650 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
1651 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
1652 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
1653 } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
1654 self._sort_formats(formats)
1659 'description': description,
1660 'thumbnail': thumbnail,
1661 'duration': duration,
1666 def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
1667 res = self._download_webpage_handle(
1669 note=note or 'Downloading MPD manifest',
1670 errnote=errnote or 'Failed to download MPD manifest',
1675 mpd_base_url = base_url(urlh.geturl())
1677 return self._parse_mpd_formats(
1678 compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url,
1679 formats_dict=formats_dict, mpd_url=mpd_url)
1681 def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
1683 Parse formats from MPD manifest.
1685 1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
1686 http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
1687 2. https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP
1689 if mpd_doc.get('type') == 'dynamic':
1692 namespace = self._search_regex(r'(?i)^{([^}]+)?}MPD$', mpd_doc.tag, 'namespace', default=None)
1695 return self._xpath_ns(path, namespace)
1697 def is_drm_protected(element):
1698 return element.find(_add_ns('ContentProtection')) is not None
1700 def extract_multisegment_info(element, ms_parent_info):
1701 ms_info = ms_parent_info.copy()
1703 # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
1704 # common attributes and elements. We will only extract relevant
1706 def extract_common(source):
1707 segment_timeline = source.find(_add_ns('SegmentTimeline'))
1708 if segment_timeline is not None:
1709 s_e = segment_timeline.findall(_add_ns('S'))
1711 ms_info['total_number'] = 0
1714 r = int(s.get('r', 0))
1715 ms_info['total_number'] += 1 + r
1716 ms_info['s'].append({
1717 't': int(s.get('t', 0)),
1718 # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
1719 'd': int(s.attrib['d']),
1722 start_number = source.get('startNumber')
1724 ms_info['start_number'] = int(start_number)
1725 timescale = source.get('timescale')
1727 ms_info['timescale'] = int(timescale)
1728 segment_duration = source.get('duration')
1729 if segment_duration:
1730 ms_info['segment_duration'] = int(segment_duration)
1732 def extract_Initialization(source):
1733 initialization = source.find(_add_ns('Initialization'))
1734 if initialization is not None:
1735 ms_info['initialization_url'] = initialization.attrib['sourceURL']
1737 segment_list = element.find(_add_ns('SegmentList'))
1738 if segment_list is not None:
1739 extract_common(segment_list)
1740 extract_Initialization(segment_list)
1741 segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
1743 ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
1745 segment_template = element.find(_add_ns('SegmentTemplate'))
1746 if segment_template is not None:
1747 extract_common(segment_template)
1748 media = segment_template.get('media')
1750 ms_info['media'] = media
1751 initialization = segment_template.get('initialization')
1753 ms_info['initialization'] = initialization
1755 extract_Initialization(segment_template)
1758 mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
1760 for period in mpd_doc.findall(_add_ns('Period')):
1761 period_duration = parse_duration(period.get('duration')) or mpd_duration
1762 period_ms_info = extract_multisegment_info(period, {
1766 for adaptation_set in period.findall(_add_ns('AdaptationSet')):
1767 if is_drm_protected(adaptation_set):
1769 adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
1770 for representation in adaptation_set.findall(_add_ns('Representation')):
1771 if is_drm_protected(representation):
1773 representation_attrib = adaptation_set.attrib.copy()
1774 representation_attrib.update(representation.attrib)
1775 # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
1776 mime_type = representation_attrib['mimeType']
1777 content_type = mime_type.split('/')[0]
1778 if content_type == 'text':
1779 # TODO implement WebVTT downloading
1781 elif content_type == 'video' or content_type == 'audio':
1783 for element in (representation, adaptation_set, period, mpd_doc):
1784 base_url_e = element.find(_add_ns('BaseURL'))
1785 if base_url_e is not None:
1786 base_url = base_url_e.text + base_url
1787 if re.match(r'^https?://', base_url):
1789 if mpd_base_url and not re.match(r'^https?://', base_url):
1790 if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
1792 base_url = mpd_base_url + base_url
1793 representation_id = representation_attrib.get('id')
1794 lang = representation_attrib.get('lang')
1795 url_el = representation.find(_add_ns('BaseURL'))
1796 filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
1797 bandwidth = int_or_none(representation_attrib.get('bandwidth'))
1799 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
1801 'manifest_url': mpd_url,
1802 'ext': mimetype2ext(mime_type),
1803 'width': int_or_none(representation_attrib.get('width')),
1804 'height': int_or_none(representation_attrib.get('height')),
1805 'tbr': int_or_none(bandwidth, 1000),
1806 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
1807 'fps': int_or_none(representation_attrib.get('frameRate')),
1808 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
1809 'format_note': 'DASH %s' % content_type,
1810 'filesize': filesize,
1812 f.update(parse_codecs(representation_attrib.get('codecs')))
1813 representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
1815 def prepare_template(template_name, identifiers):
1816 t = representation_ms_info[template_name]
1817 t = t.replace('$RepresentationID$', representation_id)
1818 t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
1819 t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
1820 t.replace('$$', '$')
1823 # @initialization is a regular template like @media one
1824 # so it should be handled just the same way (see
1825 # https://github.com/rg3/youtube-dl/issues/11605)
1826 if 'initialization' in representation_ms_info:
1827 initialization_template = prepare_template(
1829 # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
1830 # $Time$ shall not be included for @initialization thus
1831 # only $Bandwidth$ remains
1833 representation_ms_info['initialization_url'] = initialization_template % {
1834 'Bandwidth': bandwidth,
1837 if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
1839 media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
1841 # As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
1842 # can't be used at the same time
1843 if '%(Number' in media_template and 's' not in representation_ms_info:
1844 segment_duration = None
1845 if 'total_number' not in representation_ms_info and 'segment_duration':
1846 segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
1847 representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
1848 representation_ms_info['fragments'] = [{
1849 'url': media_template % {
1850 'Number': segment_number,
1851 'Bandwidth': bandwidth,
1853 'duration': segment_duration,
1854 } for segment_number in range(
1855 representation_ms_info['start_number'],
1856 representation_ms_info['total_number'] + representation_ms_info['start_number'])]
1858 # $Number*$ or $Time$ in media template with S list available
1859 # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
1860 # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
1861 representation_ms_info['fragments'] = []
1864 segment_number = representation_ms_info['start_number']
1866 def add_segment_url():
1867 segment_url = media_template % {
1868 'Time': segment_time,
1869 'Bandwidth': bandwidth,
1870 'Number': segment_number,
1872 representation_ms_info['fragments'].append({
1874 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
1877 for num, s in enumerate(representation_ms_info['s']):
1878 segment_time = s.get('t') or segment_time
1882 for r in range(s.get('r', 0)):
1883 segment_time += segment_d
1886 segment_time += segment_d
1887 elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
1889 # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
1890 # or any YouTube dashsegments video
1893 timescale = representation_ms_info['timescale']
1894 for s in representation_ms_info['s']:
1895 duration = float_or_none(s['d'], timescale)
1896 for r in range(s.get('r', 0) + 1):
1898 'url': representation_ms_info['segment_urls'][segment_index],
1899 'duration': duration,
1902 representation_ms_info['fragments'] = fragments
1903 # NB: MPD manifest may contain direct URLs to unfragmented media.
1904 # No fragments key is present in this case.
1905 if 'fragments' in representation_ms_info:
1908 'protocol': 'http_dash_segments',
1910 if 'initialization_url' in representation_ms_info:
1911 initialization_url = representation_ms_info['initialization_url']
1912 if not f.get('url'):
1913 f['url'] = initialization_url
1914 f['fragments'].append({'url': initialization_url})
1915 f['fragments'].extend(representation_ms_info['fragments'])
1916 for fragment in f['fragments']:
1917 fragment['url'] = urljoin(base_url, fragment['url'])
1919 existing_format = next(
1920 fo for fo in formats
1921 if fo['format_id'] == representation_id)
1922 except StopIteration:
1923 full_info = formats_dict.get(representation_id, {}).copy()
1925 formats.append(full_info)
1927 existing_format.update(f)
1929 self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
1932 def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
1933 res = self._download_webpage_handle(
1935 note=note or 'Downloading ISM manifest',
1936 errnote=errnote or 'Failed to download ISM manifest',
1942 return self._parse_ism_formats(
1943 compat_etree_fromstring(ism.encode('utf-8')), urlh.geturl(), ism_id)
1945 def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
1946 if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
1949 duration = int(ism_doc.attrib['Duration'])
1950 timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
1953 for stream in ism_doc.findall('StreamIndex'):
1954 stream_type = stream.get('Type')
1955 if stream_type not in ('video', 'audio'):
1957 url_pattern = stream.attrib['Url']
1958 stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
1959 stream_name = stream.get('Name')
1960 for track in stream.findall('QualityLevel'):
1961 fourcc = track.get('FourCC')
1962 # TODO: add support for WVC1 and WMAP
1963 if fourcc not in ('H264', 'AVC1', 'AACL'):
1964 self.report_warning('%s is not a supported codec' % fourcc)
1966 tbr = int(track.attrib['Bitrate']) // 1000
1967 width = int_or_none(track.get('MaxWidth'))
1968 height = int_or_none(track.get('MaxHeight'))
1969 sampling_rate = int_or_none(track.get('SamplingRate'))
1971 track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
1972 track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
1978 stream_fragments = stream.findall('c')
1979 for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
1980 fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
1981 fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
1982 fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
1983 if not fragment_ctx['duration']:
1985 next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
1987 next_fragment_time = duration
1988 fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
1989 for _ in range(fragment_repeat):
1991 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
1992 'duration': fragment_ctx['duration'] / stream_timescale,
1994 fragment_ctx['time'] += fragment_ctx['duration']
1998 format_id.append(ism_id)
2000 format_id.append(stream_name)
2001 format_id.append(compat_str(tbr))
2004 'format_id': '-'.join(format_id),
2006 'manifest_url': ism_url,
2007 'ext': 'ismv' if stream_type == 'video' else 'isma',
2011 'asr': sampling_rate,
2012 'vcodec': 'none' if stream_type == 'audio' else fourcc,
2013 'acodec': 'none' if stream_type == 'video' else fourcc,
2015 'fragments': fragments,
2016 '_download_params': {
2017 'duration': duration,
2018 'timescale': stream_timescale,
2019 'width': width or 0,
2020 'height': height or 0,
2022 'codec_private_data': track.get('CodecPrivateData'),
2023 'sampling_rate': sampling_rate,
2024 'channels': int_or_none(track.get('Channels', 2)),
2025 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
2026 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
2031 def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
2032 def absolute_url(video_url):
2033 return compat_urlparse.urljoin(base_url, video_url)
2035 def parse_content_type(content_type):
2036 if not content_type:
2038 ctr = re.search(r'(?P<mimetype>[^/]+/[^;]+)(?:;\s*codecs="?(?P<codecs>[^"]+))?', content_type)
2040 mimetype, codecs = ctr.groups()
2041 f = parse_codecs(codecs)
2042 f['ext'] = mimetype2ext(mimetype)
2046 def _media_formats(src, cur_media_type):
2047 full_url = absolute_url(src)
2048 ext = determine_ext(full_url)
2050 is_plain_url = False
2051 formats = self._extract_m3u8_formats(
2052 full_url, video_id, ext='mp4',
2053 entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id,
2054 preference=preference)
2056 is_plain_url = False
2057 formats = self._extract_mpd_formats(
2058 full_url, video_id, mpd_id=mpd_id)
2063 'vcodec': 'none' if cur_media_type == 'audio' else None,
2065 return is_plain_url, formats
2068 media_tags = [(media_tag, media_type, '')
2069 for media_tag, media_type
2070 in re.findall(r'(?s)(<(video|audio)[^>]*/>)', webpage)]
2071 media_tags.extend(re.findall(
2072 # We only allow video|audio followed by a whitespace or '>'.
2073 # Allowing more characters may end up in significant slow down (see
2074 # https://github.com/rg3/youtube-dl/issues/11979, example URL:
2075 # http://www.porntrex.com/maps/videositemap.xml).
2076 r'(?s)(<(?P<tag>video|audio)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
2077 for media_tag, media_type, media_content in media_tags:
2082 media_attributes = extract_attributes(media_tag)
2083 src = media_attributes.get('src')
2085 _, formats = _media_formats(src, media_type)
2086 media_info['formats'].extend(formats)
2087 media_info['thumbnail'] = media_attributes.get('poster')
2089 for source_tag in re.findall(r'<source[^>]+>', media_content):
2090 source_attributes = extract_attributes(source_tag)
2091 src = source_attributes.get('src')
2094 is_plain_url, formats = _media_formats(src, media_type)
2096 f = parse_content_type(source_attributes.get('type'))
2097 f.update(formats[0])
2098 media_info['formats'].append(f)
2100 media_info['formats'].extend(formats)
2101 for track_tag in re.findall(r'<track[^>]+>', media_content):
2102 track_attributes = extract_attributes(track_tag)
2103 kind = track_attributes.get('kind')
2104 if not kind or kind in ('subtitles', 'captions'):
2105 src = track_attributes.get('src')
2108 lang = track_attributes.get('srclang') or track_attributes.get('lang') or track_attributes.get('label')
2109 media_info['subtitles'].setdefault(lang, []).append({
2110 'url': absolute_url(src),
2112 if media_info['formats'] or media_info['subtitles']:
2113 entries.append(media_info)
2116 def _extract_akamai_formats(self, manifest_url, video_id, hosts={}):
2118 hdcore_sign = 'hdcore=3.7.0'
2119 f4m_url = re.sub(r'(https?://[^/+])/i/', r'\1/z/', manifest_url).replace('/master.m3u8', '/manifest.f4m')
2120 hds_host = hosts.get('hds')
2122 f4m_url = re.sub(r'(https?://)[^/]+', r'\1' + hds_host, f4m_url)
2123 if 'hdcore=' not in f4m_url:
2124 f4m_url += ('&' if '?' in f4m_url else '?') + hdcore_sign
2125 f4m_formats = self._extract_f4m_formats(
2126 f4m_url, video_id, f4m_id='hds', fatal=False)
2127 for entry in f4m_formats:
2128 entry.update({'extra_param_to_segment_url': hdcore_sign})
2129 formats.extend(f4m_formats)
2130 m3u8_url = re.sub(r'(https?://[^/]+)/z/', r'\1/i/', manifest_url).replace('/manifest.f4m', '/master.m3u8')
2131 hls_host = hosts.get('hls')
2133 m3u8_url = re.sub(r'(https?://)[^/]+', r'\1' + hls_host, m3u8_url)
2134 formats.extend(self._extract_m3u8_formats(
2135 m3u8_url, video_id, 'mp4', 'm3u8_native',
2136 m3u8_id='hls', fatal=False))
2139 def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
2140 url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
2141 url_base = self._search_regex(r'(?:https?|rtmp|rtsp)(://[^?]+)', url, 'format url')
2142 http_base_url = 'http' + url_base
2144 if 'm3u8' not in skip_protocols:
2145 formats.extend(self._extract_m3u8_formats(
2146 http_base_url + '/playlist.m3u8', video_id, 'mp4',
2147 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
2148 if 'f4m' not in skip_protocols:
2149 formats.extend(self._extract_f4m_formats(
2150 http_base_url + '/manifest.f4m',
2151 video_id, f4m_id='hds', fatal=False))
2152 if 'dash' not in skip_protocols:
2153 formats.extend(self._extract_mpd_formats(
2154 http_base_url + '/manifest.mpd',
2155 video_id, mpd_id='dash', fatal=False))
2156 if re.search(r'(?:/smil:|\.smil)', url_base):
2157 if 'smil' not in skip_protocols:
2158 rtmp_formats = self._extract_smil_formats(
2159 http_base_url + '/jwplayer.smil',
2160 video_id, fatal=False)
2161 for rtmp_format in rtmp_formats:
2162 rtsp_format = rtmp_format.copy()
2163 rtsp_format['url'] = '%s/%s' % (rtmp_format['url'], rtmp_format['play_path'])
2164 del rtsp_format['play_path']
2165 del rtsp_format['ext']
2166 rtsp_format.update({
2167 'url': rtsp_format['url'].replace('rtmp://', 'rtsp://'),
2168 'format_id': rtmp_format['format_id'].replace('rtmp', 'rtsp'),
2171 formats.extend([rtmp_format, rtsp_format])
2173 for protocol in ('rtmp', 'rtsp'):
2174 if protocol not in skip_protocols:
2176 'url': protocol + url_base,
2177 'format_id': protocol,
2178 'protocol': protocol,
2182 def _find_jwplayer_data(self, webpage, video_id=None, transform_source=js_to_json):
2184 r'jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)\.setup\s*\((?P<options>[^)]+)\)',
2188 jwplayer_data = self._parse_json(mobj.group('options'),
2190 transform_source=transform_source)
2191 except ExtractorError:
2194 if isinstance(jwplayer_data, dict):
2195 return jwplayer_data
2197 def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
2198 jwplayer_data = self._find_jwplayer_data(
2199 webpage, video_id, transform_source=js_to_json)
2200 return self._parse_jwplayer_data(
2201 jwplayer_data, video_id, *args, **kwargs)
2203 def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
2204 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2205 # JWPlayer backward compatibility: flattened playlists
2206 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
2207 if 'playlist' not in jwplayer_data:
2208 jwplayer_data = {'playlist': [jwplayer_data]}
2212 # JWPlayer backward compatibility: single playlist item
2213 # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
2214 if not isinstance(jwplayer_data['playlist'], list):
2215 jwplayer_data['playlist'] = [jwplayer_data['playlist']]
2217 for video_data in jwplayer_data['playlist']:
2218 # JWPlayer backward compatibility: flattened sources
2219 # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
2220 if 'sources' not in video_data:
2221 video_data['sources'] = [video_data]
2223 this_video_id = video_id or video_data['mediaid']
2225 formats = self._parse_jwplayer_formats(
2226 video_data['sources'], video_id=this_video_id, m3u8_id=m3u8_id,
2227 mpd_id=mpd_id, rtmp_params=rtmp_params, base_url=base_url)
2228 self._sort_formats(formats)
2231 tracks = video_data.get('tracks')
2232 if tracks and isinstance(tracks, list):
2233 for track in tracks:
2234 if track.get('kind') != 'captions':
2236 track_url = urljoin(base_url, track.get('file'))
2239 subtitles.setdefault(track.get('label') or 'en', []).append({
2240 'url': self._proto_relative_url(track_url)
2244 'id': this_video_id,
2245 'title': video_data['title'] if require_title else video_data.get('title'),
2246 'description': video_data.get('description'),
2247 'thumbnail': self._proto_relative_url(video_data.get('image')),
2248 'timestamp': int_or_none(video_data.get('pubdate')),
2249 'duration': float_or_none(jwplayer_data.get('duration') or video_data.get('duration')),
2250 'subtitles': subtitles,
2253 if len(entries) == 1:
2256 return self.playlist_result(entries)
2258 def _parse_jwplayer_formats(self, jwplayer_sources_data, video_id=None,
2259 m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
2261 for source in jwplayer_sources_data:
2262 source_url = self._proto_relative_url(source['file'])
2264 source_url = compat_urlparse.urljoin(base_url, source_url)
2265 source_type = source.get('type') or ''
2266 ext = mimetype2ext(source_type) or determine_ext(source_url)
2267 if source_type == 'hls' or ext == 'm3u8':
2268 formats.extend(self._extract_m3u8_formats(
2269 source_url, video_id, 'mp4', entry_protocol='m3u8_native',
2270 m3u8_id=m3u8_id, fatal=False))
2272 formats.extend(self._extract_mpd_formats(
2273 source_url, video_id, mpd_id=mpd_id, fatal=False))
2275 formats.extend(self._extract_smil_formats(
2276 source_url, video_id, fatal=False))
2277 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
2278 elif source_type.startswith('audio') or ext in (
2279 'oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
2286 height = int_or_none(source.get('height'))
2288 # Often no height is provided but there is a label in
2289 # format like "1080p", "720p SD", or 1080.
2290 height = int_or_none(self._search_regex(
2291 r'^(\d{3,4})[pP]?(?:\b|$)', compat_str(source.get('label') or ''),
2292 'height', default=None))
2295 'width': int_or_none(source.get('width')),
2297 'tbr': int_or_none(source.get('bitrate')),
2300 if source_url.startswith('rtmp'):
2301 a_format['ext'] = 'flv'
2302 # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
2303 # of jwplayer.flash.swf
2304 rtmp_url_parts = re.split(
2305 r'((?:mp4|mp3|flv):)', source_url, 1)
2306 if len(rtmp_url_parts) == 3:
2307 rtmp_url, prefix, play_path = rtmp_url_parts
2310 'play_path': prefix + play_path,
2313 a_format.update(rtmp_params)
2314 formats.append(a_format)
2317 def _live_title(self, name):
2318 """ Generate the title for a live video """
2319 now = datetime.datetime.now()
2320 now_str = now.strftime('%Y-%m-%d %H:%M')
2321 return name + ' ' + now_str
2323 def _int(self, v, name, fatal=False, **kwargs):
2324 res = int_or_none(v, **kwargs)
2325 if 'get_attr' in kwargs:
2326 print(getattr(v, kwargs['get_attr']))
2328 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2330 raise ExtractorError(msg)
2332 self._downloader.report_warning(msg)
2335 def _float(self, v, name, fatal=False, **kwargs):
2336 res = float_or_none(v, **kwargs)
2338 msg = 'Failed to extract %s: Could not parse value %r' % (name, v)
2340 raise ExtractorError(msg)
2342 self._downloader.report_warning(msg)
2345 def _set_cookie(self, domain, name, value, expire_time=None):
2346 cookie = compat_cookiejar.Cookie(
2347 0, name, value, None, None, domain, None,
2348 None, '/', True, False, expire_time, '', None, None, None)
2349 self._downloader.cookiejar.set_cookie(cookie)
2351 def _get_cookies(self, url):
2352 """ Return a compat_cookies.SimpleCookie with the cookies for the url """
2353 req = sanitized_Request(url)
2354 self._downloader.cookiejar.add_cookie_header(req)
2355 return compat_cookies.SimpleCookie(req.get_header('Cookie'))
2357 def get_testcases(self, include_onlymatching=False):
2358 t = getattr(self, '_TEST', None)
2360 assert not hasattr(self, '_TESTS'), \
2361 '%s has _TEST and _TESTS' % type(self).__name__
2364 tests = getattr(self, '_TESTS', [])
2366 if not include_onlymatching and t.get('only_matching', False):
2368 t['name'] = type(self).__name__[:-len('IE')]
2371 def is_suitable(self, age_limit):
2372 """ Test whether the extractor is generally suitable for the given
2373 age limit (i.e. pornographic sites are not, all others usually are) """
2375 any_restricted = False
2376 for tc in self.get_testcases(include_onlymatching=False):
2377 if tc.get('playlist', []):
2378 tc = tc['playlist'][0]
2379 is_restricted = age_restricted(
2380 tc.get('info_dict', {}).get('age_limit'), age_limit)
2381 if not is_restricted:
2383 any_restricted = any_restricted or is_restricted
2384 return not any_restricted
2386 def extract_subtitles(self, *args, **kwargs):
2387 if (self._downloader.params.get('writesubtitles', False) or
2388 self._downloader.params.get('listsubtitles')):
2389 return self._get_subtitles(*args, **kwargs)
2392 def _get_subtitles(self, *args, **kwargs):
2393 raise NotImplementedError('This method must be implemented by subclasses')
2396 def _merge_subtitle_items(subtitle_list1, subtitle_list2):
2397 """ Merge subtitle items for one language. Items with duplicated URLs
2398 will be dropped. """
2399 list1_urls = set([item['url'] for item in subtitle_list1])
2400 ret = list(subtitle_list1)
2401 ret.extend([item for item in subtitle_list2 if item['url'] not in list1_urls])
2405 def _merge_subtitles(cls, subtitle_dict1, subtitle_dict2):
2406 """ Merge two subtitle dictionaries, language by language. """
2407 ret = dict(subtitle_dict1)
2408 for lang in subtitle_dict2:
2409 ret[lang] = cls._merge_subtitle_items(subtitle_dict1.get(lang, []), subtitle_dict2[lang])
2412 def extract_automatic_captions(self, *args, **kwargs):
2413 if (self._downloader.params.get('writeautomaticsub', False) or
2414 self._downloader.params.get('listsubtitles')):
2415 return self._get_automatic_captions(*args, **kwargs)
2418 def _get_automatic_captions(self, *args, **kwargs):
2419 raise NotImplementedError('This method must be implemented by subclasses')
2421 def mark_watched(self, *args, **kwargs):
2422 if (self._downloader.params.get('mark_watched', False) and
2423 (self._get_login_info()[0] is not None or
2424 self._downloader.params.get('cookiefile') is not None)):
2425 self._mark_watched(*args, **kwargs)
2427 def _mark_watched(self, *args, **kwargs):
2428 raise NotImplementedError('This method must be implemented by subclasses')
2430 def geo_verification_headers(self):
2432 geo_verification_proxy = self._downloader.params.get('geo_verification_proxy')
2433 if geo_verification_proxy:
2434 headers['Ytdl-request-proxy'] = geo_verification_proxy
2437 def _generic_id(self, url):
2438 return compat_urllib_parse_unquote(os.path.splitext(url.rstrip('/').split('/')[-1])[0])
2440 def _generic_title(self, url):
2441 return compat_urllib_parse_unquote(os.path.splitext(url_basename(url))[0])
2444 class SearchInfoExtractor(InfoExtractor):
2446 Base class for paged search queries extractors.
2447 They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query}
2448 Instances should define _SEARCH_KEY and _MAX_RESULTS.
2452 def _make_valid_url(cls):
2453 return r'%s(?P<prefix>|[1-9][0-9]*|all):(?P<query>[\s\S]+)' % cls._SEARCH_KEY
2456 def suitable(cls, url):
2457 return re.match(cls._make_valid_url(), url) is not None
2459 def _real_extract(self, query):
2460 mobj = re.match(self._make_valid_url(), query)
2462 raise ExtractorError('Invalid search query "%s"' % query)
2464 prefix = mobj.group('prefix')
2465 query = mobj.group('query')
2467 return self._get_n_results(query, 1)
2468 elif prefix == 'all':
2469 return self._get_n_results(query, self._MAX_RESULTS)
2473 raise ExtractorError('invalid download number %s for query "%s"' % (n, query))
2474 elif n > self._MAX_RESULTS:
2475 self._downloader.report_warning('%s returns max %i results (you requested %i)' % (self._SEARCH_KEY, self._MAX_RESULTS, n))
2476 n = self._MAX_RESULTS
2477 return self._get_n_results(query, n)
2479 def _get_n_results(self, query, n):
2480 """Get a specified number of results for a query"""
2481 raise NotImplementedError('This method must be implemented by subclasses')
2484 def SEARCH_KEY(self):
2485 return self._SEARCH_KEY